Snowflake’s annual user conference is under way, so we have a blizzard (sorry not sorry) of announcements from and about them. I’ll just highlight private previews of a new Large Language Model, based on last year’s Applica acquisition. It will focus on making unstructured data more usable, and thus encouraging companies to store more of it in Snowflake.
RudderStack has launched a feature to build unified customer profiles in a Snowflake data warehouse. The new feature, called, um, Profiles, builds an identity graph and uses it to assemble a customer data table that can include user-defined attributes such as lifetime purchases and last interaction date. It joins Rudderstack’s existing tools for data collection and distribution to create what’s effectively an integrated CDP. The main difference from a conventional CDP is the same Snowflake instance could contain non-CDP data.
Alliances between ad platforms and retail sales measurement vendors have become so common I may stop writing about them. They were originally worth noting because they’re one way that marketers adjust to the loss of data from third-party cookies. Before dropping the veil of silence, I’ll mention that self-serve programmatic ad platform StackAdapt now gives its clients access to retail foot traffic data from Precisely PlaceIQ.
After years of thinking they were in the business of selling things to people, retailers have realized they’re actually in the business of selling people to advertisers. Retail media networks are the most obvious result, but why stop at ads sold directly by the retailer? Walmart has hooked up with Disney Advertising to test letting advertisers on Disney+, Hulu, and other Disney outlets use Walmart audience information to improve targeting and measure results.